Skip to content

🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.

update rlaif

jingyaogong4a7c1c49e8ea
diff --git a/model/dataset.py b/model/dataset.py
index 4336c7e..d67cb8c 100644
--- a/model/dataset.py
+++ b/model/dataset.py
@@ -196,5 +196,50 @@ class DPODataset(Dataset):
return loss_mask
+class RLAIFDataset(Dataset):
+ def __init__(self, jsonl_path, tokenizer, max_length=1024):
+ super().__init__()
+ self.tokenizer = tokenizer
+ self.max_length = max_length
+ self.samples = self.load_data(jsonl_path)
+ self.bos_id = tokenizer('<s>assistant', add_special_tokens=False).input_ids
+ self.eos_id = tokenizer('</s>', add_special_tokens=False).input_ids
+
+ def __len__(self):
+ return len(self.samples)
+
+ def load_data(self, path):
+ samples = []
+ with open(path, 'r', encoding='utf-8') as f:
+ for line_num, line in enumerate(f, 1):
+ data = json.loads(line.strip())
+ samples.append(data)
+ return samples
+
+ def _create_chat_prompt(self, conversations):
+ """构建符合ChatML格式的对话"""
+ messages = []
+ answer = ''
+ for i, turn in enumerate(conversations):
+ role = 'user' if i % 2 == 0 else 'assistant'
+ messages.append({"role": role, "content": turn['content']})
+ answer = turn['content']
+ return self.tokenizer.apply_chat_template(
+ messages[:-1],
+ tokenize=False,
+ add_generation_prompt=True
+ ), answer
+
+ def __getitem__(self, index):
+ sample = self.samples[index]
+ # 构建对话提示
+ prompt, answer = self._create_chat_prompt(sample['conversations'])
+
+ return {
+ 'prompt': prompt,
+ 'answer': answer
+ }
+
+
if __name__ == "__main__":
pass